Experts flag methodology concerns in unauthored seven-page report
A seven-page, unauthored Census Bureau analysis on noncitizen voting in the 2020 election that President Trump is using to claim widespread illegal voting was prepared in part by individuals affiliated with the America First Policy Institute, a Trump-aligned think tank, according to NPR. Election and data experts are questioning the report’s methodology and its place among disputed federal figures on noncitizen voting ahead of the fall midterm elections.
The report, titled “Noncitizen Voting in the 2020 Election, A Beginning Analysis,” draws on a 2020 commercial voter file and various federal administrative records, according to its text. It says the bureau examined more than 128 million voter records and identified approximately 24,000 noncitizens among voters who cast ballots in the 2020 election. The bureau has yet to check an additional 32 million records, the report says.
In a social media post Tuesday, Trump said the initial findings “prove that over 24,000 Noncitizens voted illegally” in the 2020 presidential election and that the final number would “explode” once the statistical agency completed its analysis of the remaining records. NPR reports that Trump continues to falsely claim the 2020 election was rigged; in the final stretch before this fall’s midterm elections, Trump has urged his Republican allies in Congress to pass sweeping new voting restrictions.
Work on the analysis was not conducted by career civil servants at the Census Bureau, according to a bureau employee who NPR agreed not to name because they fear retaliation at work. The team behind it included individuals affiliated with the America First Policy Institute, a think tank started by officials from Trump’s first administration. Spokespeople for the bureau, the AFPI and the White House did not respond to NPR’s requests for comment.
AFPI highlighted the analysis in a social media post Tuesday, saying the think tank “maintains that maximizing transparency is a critical step to restoring confidence in our elections.”
Election experts pushed back against the report’s findings. David Becker, executive director of the Center for Election Innovation & Research, said in a text exchange with NPR: “This analysis is laughable. Census data cannot reliably be matched to voter data. That said, it’s good to see the president admit these numbers are so small that they couldn’t have changed the outcome in 2020.”
The report’s own numbers, scaled to the records analyzed, are proportionally small. Some 24,000 votes out of 128 million is just shy of 0.02%.
Michael Morse, an assistant professor at the University of Pennsylvania law school who has researched voter registration files, wrote in an email to NPR: “I’d expect that the Census data matching [exercise] is plagued by the lack of unique identifiers in voter records and outdated or incomplete naturalization records. But the brief’s lack of transparency makes it difficult to pin down each problem or trust any explanation. Regardless, the estimated rate of apparent non-citizen voting is both a tiny fraction of total votes … and entirely consistent with both types of error.”
Former Census Bureau employees raised separate concerns about the report’s structure. Misty Heggeness, a former bureau economist whose work is cited in the report, said: “My biggest concern would be there’s no author attached to the report, which is not common practice within the Census Bureau. Usually there are layers and layers of rigor and testing because the Census Bureau is a neutral statistical agency. They want to make sure that they’re getting it right and understanding the data. The fact that there’s no author on this report leads me to wonder whether or not that did happen.”
Amy O’Hara, a former chief of the bureau’s Center for Administrative Records Research and Applications who now directs Georgetown University’s Massive Data Institute, highlighted the risk that mismatches between voter files and citizenship records from other federal agencies could inflate the count. “This brief lacks any sort of measures of variance or discussion of bias,” O’Hara said. “While those matches were found, were they the same people between the voter file and the compilation of data from Homeland Security, or do you have people with the same names potentially living in the same state and representing a mismatch?”
O’Hara also questioned the report’s claim that the bureau is “uniquely capable of compiling and publishing statistics on election crime and other forms of fraud.” “I found that very surprising because even if they observed anomalies in the data, those anomalies aren’t necessarily fraud,” she said.
The report’s release follows a major exodus of statisticians, economists and other career civil servants from the bureau after the Trump administration’s workforce cuts. The administration has signaled it is preparing to propose major changes to the 2030 census that are likely to put the bureau at the center of another political and legal fight.
O’Hara framed the report as another threat to the bureau’s reputation as a nonpartisan statistical agency. “We are drawing in on the 2030 census that really requires the participation of everybody in the country. Anything that casts a negative light on the Census Bureau’s motives or operations or transparency or lack of adherence to policy could be detrimental,” she said.
The figures from the Census Bureau report join a series the Trump administration has released in recent months in an attempt to quantify noncitizen voting in federal races, a phenomenon that election audits and academic studies have shown to be rare. Last month, the Department of Homeland Security announced it had found 250,000 noncitizens on public voter rolls in California, Nevada, New Jersey and Pennsylvania, but has yet to publicly disclose how it made those calculations. Election officials in those states have expressed doubts about the accuracy of those claims.
Heggeness, now a professor of public affairs and economics at the University of Kansas, said the public should treat the report with caution given the absence of standard scholarly apparatus. “This document really does have a lack of a source and citations and explanation about methods,” she said. “And if the public sees a document like this that lacks all of those really integral pieces that are generally put out with official statistics, that’s a really good sign to question the validity of the document.”